Cytokines released during infection/inflammation can act on the brain to activate the hypothalamo-pituitary-adrenal (HPA) axis. Glucocorticoid mediators of HPA function serve critical metabolic and immunoregulatory roles that facilitate coping with the immediate threat, and can also impact inflammatory disease processes. Work on this project has defined a model for immune-mediated HPA engagement that involves (1) binding of circulating cytokines by cerebrovascular endothelial cells (ECs), which (2) activate resident macrophages in perivascular spaces (perivascular cells; PVCs) to cyclooxygenase-dependent prostaglandin E2 (PGE2) release. ECs' capacity to display inducible PGE2 synthesis is negatively regulated by PVCs. PGE2 then (3) stimulates brainstem catecholamine neurons that express PGE2 receptors, and, via their axonal projections, (4) hypothalamic neurosecretory neurons that initiate HPA output. We propose to test and refine this model, assess its generality and the potential of manipulating it for intervening in CNS inflammatory disease. First, to test a posited mechanism underlying potent anti-inflammatory influences of PVCs on ECs, via 15-deoxy- ?12,14-prostaglandin J2 (15d-PGJ2) binding by peroxisome proliferator-activated receptor-? (PPAR?), responses of a conditional, EC-specific PPAR? null mouse model to endotoxin challenges will be assessed at each level of the circuit. Transcriptional profiling of PVCs and ECs will be carried out to identify candidate mediators of EC stimulatory influences on PVCs, whose involvement will then be tested. Second, combined axonal transport and histochemical methods will be used to identify sources of endotoxin-sensitive cells that project to hypothalamus and express PGE2 receptors (EP1-4R). Conditional knockout models will assess the role of the EP?R in CNS in general, and in catecholamine neurons specifically, in LPS-induced HPA activation. Third, tract tracing, histochemical and ablation approaches will be used to clarify the hierarchical organization of CNS cell groups that provide for HPA recruitment by insults that differ in nature and intensity. Fourth, we will pursue evidence that variants of the model are operative in engaging CNS/HPA responses to a prototypic emotional stressor, restraint, and an alternate immune challenge presented by bacterial superantigen. Finally, we will test the involvement of the vascular 15d-PGJ2 - PPAR? anti-inflammatory signaling on clinical and neuropathological sequelae of a mouse model of multiple sclerosis, and in the exacerbation of neuropathology in an Alzheimer's Disease model by systemic inflammatory challenges.